OIG Data Brief Causes a Stir
By Monique Griffin
For The Record
Vol. 33 No. 4 P. 24
In February, the government agency said data indicate hospitals are engaging in upcoding. Citing several reasons, industry experts disagree with the assertations.
Earlier this year, the Office of Inspector General (OIG) released a data brief concerning the pre-COVID trend toward more expensive inpatient hospital stays in Medicare. OIG analyzed paid Medicare Part A claims for inpatient hospital stays from fiscal year (FY )2014 through FY 2019 to identify the trend.
OIG found hospitals are increasingly billing inpatient stays at the highest severity (most expensive). Severity levels are determined by the Medicare Severity diagnosis-related group (MS-DRG). In the statistical time period, nearly one-half of all Medicare spending on inpatient stays was at the highest severity level, which represented a 20% increase from FY 2014 to FY 2019. During the same time frame, the number of stays at the other severity levels decreased. Lower-level severity lengths of stay remained largely the same, but high-severity stays decreased in length.
OIG contends the high-severity stays are more vulnerable to upcoding (billing at a higher level than appropriate or that can be clinically validated). One-third of these stays were deemed to have lasted a “particularly short amount of time” and more than one-half of the stays had only one diagnosis that qualified the stay for payment at the highest level.
OIG recommended that the Centers for Medicare & Medicaid Services (CMS) conduct targeted reviews of MS-DRGs, stays that are vulnerable to upcoding, and the hospitals that frequently bill them.
Some health care experts disagree with the findings, sighting inconclusive data, a lack of understanding regarding patient medical management, and the role of continuing advances in treatment and medical technology.
“I found it interesting that this audit looked at admissions using FY 2014 as the baseline, which is exactly when the Two Midnight Rule started,” says Ronald Hirsch, MD, FACP, CHCQM, CHRI, vice president of regulations and education group at R1 RCM. “Seven and a half years later, the rule still confuses doctors, but certainly using 2014 as a baseline for this audit created statistical uncertainty.”
The Two Midnight Rule states that in order for inpatient admissions to be payable under Medicare Part A, the admitting physician expects the patient to require a hospital stay that crosses two midnights and the medical necessity is supported in the patient record. It also allows for greater flexibility for determining when an admission not lasting two midnights should be payable under Part A. These exceptions would be decided on a case-by-case basis hinging on the documentation in the patient’s medical record and subject to medical review.
“In addition to the confusion, with time doctors became more familiar with the rule,” Hirsch says. “The lower-acuity patients who were less likely to have CCs [complications or comorbidities] or MCCs [major complications or comorbidities] declined as they were treated as outpatients with observation services, leaving the inpatients with higher acuity and a higher percentage with CCs and MCCs. It's what the Two Midnight Rule was designed to do.”
Hirsch continues, “In addition, the OIG made a significant error. They repeatedly state, ‘Medicare considers each secondary diagnosis to be a major complication, a minor complication, or not a complication.’ As we know, they are called CC and MCC because the CC stands for comorbidity and complication. If looking at the lens of it being simply ‘complication,’ I can absolutely understand why the OIG would be worried that there are significantly more complications and patients with complications should have longer stays, not shorter ones.
“But the reality is that the incidence of comorbid conditions that are captured in documentation and coding, which CMS acknowledged is appropriate, is climbing and that likely accounts for the changes rather than an increasing number of complications.”
William E. Haik, MD, FCCP, CDIP, a director at DRG Review, questions the data used by OIG in the data brief. “The obvious problem with the OIG review is that it’s done through Medicare Part A data,” he says. “They didn’t actually review any health records. This is important because with Medicare Part A, health records are not clinically validated, unlike audits by Medicare Part C or Medicare Advantage, and therefore there would be more CCs and MCCs reported.
“For Medicare Part A, the RACs [recovery audit contractors] are only allowed to look at the coding and the chart documentation. They kind of dip their toes into the use of the ‘General Rule’ when it applies to the reporting of any additional diagnoses.”
Haik goes further: “So, really, they’re doing a 20,000-foot audit of cases that are not clinically reviewed. It’s like a double error as far as I’m concerned. That’s one problem.
“The second problem refers to all statistical analyses … there’s lies, there’s damn lies, and then there are statistics. You always have to question these kinds of statistical reviews without any kind of clinical validation.”
In a memo response from Seema Verma (now past CMS Administrator), CMS specifically stated that in absence of medical record reviews, it is not possible to determine whether higher-severity stays are a result of upcoding as opposed to other factors such as hospitals with specialty services that would logically draw more patients with CCs and MCCs.
Haik also sees this correlation. “They [OIG] drill down to these 5% of hospitals that have a large percentage of the sicker folks. I can imagine that would be common. I think that would not be unusual since a lot of rural hospitals are transferring their severely ill patients to tertiary hospitals.
“So I can clearly see why you would have a concentration of super sick patients in one type of hospital, which they did not break out in their study,” he says.
CMS Chimes In
CMS also defended its rigorous audit process. From the memo: “CMS uses a robust program integrity strategy to reduce and prevent Medicare improper payments, including automated system edits within the claims processing system and prepayment and postpayment medical reviews. CMS also uses the Fraud Prevention System to analyze Medicare fee-for-service claims using sophisticated algorithms to target investigative resources, generate alerts for suspect claims or providers and suppliers, and provide information to facilitate and support investigations of the most egregious, suspect, or aberrant activity.”
Importantly, CMS notes that RACs are incentivized to investigate improper coding; they are paid a contingency fee based on percentages recovered from or reimbursed to providers. Those “higher-severity inpatient stays” based on MS-DRGs would be one of the areas targeted by RACs.
Alicia M. Gordon, RN, CCDS, CCDS-O, CDIP, CCS, manager of the national reimbursement practice at CDI Services, says, “I believe CMS’s response was logical and appropriate.”
“In their response to the findings, CMS disagreed with the OIG on every point,” Hirsch says. “CMS feels they are diligent in ensuring compliance with existing rules and regulations. CMS noted that they have the RACs auditing clinical validation and the BFCC-QIOs [Beneficiary and Family Centered Care-Quality Improvement Organizations] are tasked with auditing higher-weighted DRGs.
“CMS famously said in 2007, ‘We believe it is important to address the notion in some of the public comments that CMS believes changes in how services are documented or coded that is consistent with the medical record is inappropriate or otherwise unethical. We do not believe there is anything inappropriate, unethical, or otherwise wrong with hospitals taking full advantage of coding opportunities to maximize Medicare payment that is supported by documentation in the medical record.’”
In its public comments, MedPAC (the Medicare Payment Advisory Commission) recommended an adjustment for improvements in documentation and coding and noted that hospital efforts to improve the specificity and accuracy of documentation and coding are perfectly legitimate.
“I think CMS’s response is 100% correct,” Haik says.
In response to the OIG’s concern about the more severe stays having lasted a “particularly short amount of time,” advances in medicine, documentation, and technology are the cause, experts say.
“Of course,” Hirsch says. “Physicians are also getting better at treating patients efficiently, with newer, more potent therapies. Much work has been done on ensuring patients are discharged more thoughtfully and with clear follow-up plans. So they [providers] may feel more comfortable that the patients will have proper outpatient care and be more willing to discharge the patient out of the hospital to home, an environment where the patient is more comfortable and safer.”
Haik agrees. “One reason I think we are getting people out of the hospital quicker, even though they are sicker, is we’ve had significant advances in technology,” he says, adding that a lag in updating DRGs also plays a role. “Although the DRG system is supposed to be recalibrated every year to reflect lengths of stays, it’s a slow-moving animal. It’s like turning a battleship around in a creek. You’ve got to take your time doing it. That’s one reason why I think severity and lengths of stay don’t match.”
CDI and Coding
Clinical documentation improvement (CDI) advances have increased the documenting of CCs and MCCs with additional responsibility and the weight of integrity.
“The data brief emphasizes the need for coders, coding managers, and revenue cycle leaders to continue their focus on compliant coding and ensuring appropriate, accurate coding depth. Any diagnosis coded must have the appropriate clinical indicators and be able to be clinically validated,” Gordon says.
“It is incumbent on every provider to ensure that if a diagnosis is documented, it is a clinically valid diagnosis, whether that diagnosis is a CC/MCC or not,” Hirsch says. “While the official coding guidelines may have stated, ‘The assignment of a diagnosis code is based on the provider’s diagnostic statement that the condition exists. The provider’s statement that the patient has a particular condition is sufficient. Code assignment is not based on clinical criteria used by the provider to establish the diagnosis,’ coders must also follow the Uniform Hospital Discharge Data Set guidelines that say reportable diagnoses must be evaluated, treated, diagnostically tested, used additional nursing services, or extended the length of stay.
“This is an area of opportunity for clinical documentation integrity specialists to use their expertise to ensure the documented diagnoses are not only clinically valid but also capturing the appropriate specificity and acuity for each patient.”
“Another reason we are seeing an increase in severity is because the CDI programs are becoming much more robust,” Haik says. “I think in a lot of ways, that’s a good thing. It helps hospitals and physicians to get credit for the level of severity of patients they are taking care of because of better documentation.”
Hirsch delves into coding as a whole.
“Coders must also capture all codes, whether they affect payment or not. The OIG noted an increasing number of claims with a single CC or MCC. The admission with one CC or MCC is not in itself problematic, but that admission is at higher risk of a DRG downgrade without any CCs or MCCs standing by in case one gets removed for lack of clinical validation,” he says.
“CDI should be carefully reviewing the charts with only one CC or MCC to make sure the documentation is well supported clinically and to ensure that any other potential CCs or MCCs are documented by a provider in a way that coding can capture it.”
Hirsch advocates strongly for the use of codes for the social determinants of health (SDOH), V55 to V65. “While none are designated as CCs or MCCs, their use is critical to demonstrate the effects that the SDOH have on health care,” he says. “These codes can be coded from nonphysician documentation, which is good in that the doctor does not have to learn a new naming paradigm but bad in that coders are not used to coding from nonphysician documentation. One day soon CMS will proceed, as they tried in 2019, to establish the SDOH codes as CCs and MCCs and give them weight in the hierarchal condition category system.”
While CMS and OIG may not agree on the data brief, very little may be changed by its release.
“I suspect there will be no change,” Haik says. “Unless Medicare allows the RACs to clinically validate traditional Medicare cases, I think they are already focusing on single MCC charts. It’s already there. They have their review edits. That’s a no-brainer and it’s already being done.”
“I expect the recovery audit contractors will have a focus on the listed DRGs in the report,” Gordon says. But, importantly, she adds, “Any diagnosis captured should tell the patient’s story accurately, and the selection of the most accurate principal diagnosis has never been more essential.”
Hirsch summarizes the true goal behind the data brief, rules and regulations, and other clinical medical discussions.
“In my teachings, it is all about accuracy and compliance,” he says. “While there may be additional revenue associated with improved documentation and coding, the more important outcome is improved communication between caregivers and more accurate portrayal of the acuity of a patient's illness.
“While there always will be bad actors, the vast majority of physicians and hospitals are doing their best to provide excellent care to patients and are simply adapting to the changing rules and guidelines, just the same as our medical interventions have changed over time.”
“It is just good health care,” Haik says.
— Monique Griffin is a CPC-A and a writer.